[Feature] Add dynamic dispatch import resolution across WASM entry points by marshacb · Pull Request #1208 · ProvableHQ/sdk

marshacb · 2026-02-26T05:05:30Z

Motivation

Programs using call.dynamic reference their targets via field-encoded identifiers at runtime, not through static import declarations. The existing resolve_imports only loads statically declared imports, so dynamically dispatched targets would never be added to the process which would cause authorization and execution to fail for any program using call.dynamic.

This PR:

Unifies import resolution into a single resolve_imports_or_builder dispatcher across all WASM entry points (execute, authorize, proving request, and deploy paths)
Introduces ProgramImportsBuilder, a zero-cost WASM builder that carries proving/verifying keys alongside program source, with programNames()/getProgram() for lightweight enumeration and toObject()/fromObject() for full round-trip serialization including keys as Uint8Array
Adds a KeyStore interface for persistent key storage, with automatic load-before-execution (loadKeysFromStore) and save-after-execution (persistExtractedKeys) so the second execution of a program skips key synthesis entirely
Includes IndexedDBKeyStore, a browser-native KeyStore implementation for the React template
Handles the new Literal::Identifier variant from the latest snarkVM rev

Test Plan

WASM tests (Rust):

test_resolve_dynamic_imports - verifies resolve_imports does NOT load dynamic targets, and resolve_dynamic_imports does
test_resolve_multiple_dynamic_imports - multiple dynamic targets loaded from a single imports object
test_authorize_dynamic_dispatch - process.authorize() succeeds for a call.dynamic program after import resolution
test_program_imports_program_names - lightweight enumeration without key serialization
test_program_imports_get_program - source retrieval for known/unknown programs
test_program_imports_add_key_bytes_* - byte-level key round-trip, validation of missing program, and invalid bytes
test_program_imports_to_object_from_object_roundtrip_keys - full structured format round-trip with real VK bytes

SDK tests (TypeScript):

Dynamic dispatch tests - fixture parsing, offline execution with single/multiple dynamic targets, authorization with serialization round-trip, delegated proving path
KeyStore integration tests - buildProgramImports with static/dynamic imports and network merge, loadKeysFromStore positive/negative/error paths, persistExtractedKeys with top-level keys, resolveTopLevelKeys fallback chain, synthesizeKeys auto-persist, run() end-to-end with auto-built builder and post-execution persistence

Manual testing:

Node template (template-node-ts) updated to use LocalFileKeyStore with setKeyStore() for persistent key caching on disk
React template (template-react-ts) with IndexedDBKeyStore - verified keys persist to IndexedDB after first execution, second execution logs Inserting externally provided proving and verifying keys and skips synthesis

Note

Medium Risk
Touches core WASM ProgramManager execution/authorization/deploy/proving-request paths; incorrect import loading or clone semantics could break transaction building for existing consumers.

Overview
Adds ProgramManager::resolve_dynamic_imports to load non-statically-imported programs from the user-provided imports map, enabling programs that use call.dynamic to authorize/execute successfully.

Wires this dynamic import resolution (and associated imports.clone() adjustments) into the WASM ProgramManager entry points for authorization, execution (offline/on-chain/devnode/fee estimation), deployment/upgrade, and proving requests.

Expands e2e and SDK test coverage with simple call.dynamic programs (single + multiple targets) and new tests for ProgramManager.run(), buildAuthorization/buildAuthorizationUnchecked, provingRequest, plus WASM unit tests validating dynamic import loading behavior.

^{Written by Cursor Bugbot for commit af672cd. This will update automatically on new commits. Configure here.}

marshacb · 2026-02-26T15:36:56Z

@cursor review

cursor

✅ Bugbot reviewed your changes and found no new issues!

Comment @cursor review or bugbot run to trigger another review on this PR

Copilot

Pull request overview

This PR adds support for dynamic dispatch (call.dynamic) import resolution in WASM entry points. Programs using call.dynamic reference targets via field-encoded identifiers at runtime rather than static import declarations, so the existing resolve_imports function did not load these dynamically dispatched targets. This PR introduces resolve_dynamic_imports to load all programs from the user-provided imports object that aren't already in the process, enabling authorization and execution to succeed for programs using call.dynamic.

Changes:

Adds resolve_dynamic_imports function to load all programs from imports object not already in the process
Integrates dynamic import resolution into 14 WASM entry points across execute, authorize, deploy, and proving request paths
Adds comprehensive test coverage with simple dynamic dispatch programs and multi-target scenarios

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
wasm/src/programs/manager/mod.rs	Implements `resolve_dynamic_imports` function and adds test programs and unit tests
wasm/src/programs/manager/proving_request.rs	Adds dynamic import resolution to proving request path
wasm/src/programs/manager/execute.rs	Adds dynamic import resolution to 6 execution entry points and authorization test
wasm/src/programs/manager/deploy.rs	Adds dynamic import resolution to 5 deployment/upgrade entry points
wasm/src/programs/manager/authorize.rs	Adds dynamic import resolution to authorization entry points
sdk/tests/dynamic-dispatch.test.ts	Adds SDK tests for authorization and proving request with dynamic dispatch
sdk/tests/data/dynamic-dispatch.ts	Adds test program constants and field-encoded identifiers
e2e/dynamic/index.js	Adds e2e execution tests for single and multiple dynamic imports

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

Copilot

Pull request overview

Copilot reviewed 13 out of 14 changed files in this pull request and generated 7 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

iamalwaysuncomfortable

This is a great first start, but I believe we need to add a few more things here.

Namely:

We want to add an option at the top level for people to send in program imports that include the proving and verifying keys into the top level of calls of the program manager.
We want to add the ability for keys to be resolved via any object implementing the KeyProvider interface so if keys have been stored, it can retrieve them prior to any call.
In resolve_program imports recursion we probably want to ensure we're adding in any proving and verifying keys stored.

iamalwaysuncomfortable · 2026-03-09T20:21:08Z

+#[wasm_bindgen]
+#[derive(Clone)]
+pub struct ProgramImports {
+    entries: HashMap<String, ProgramEntry>,


Suggested change

entries: HashMap<String, ProgramEntry>,

entries: HashMap<ProgramIDNative, ProgramEntry>,

We want to store the ProgramID here instead I think.

iamalwaysuncomfortable · 2026-03-09T20:33:25Z

    program?: string | Program;
    imports?: ProgramImports;
    edition?: number,
+    programImportsBuilder?: ProgramImportsBuilder;


We could allow this as an option parameter for flexibility, but MOSTLY construct keys inside the ProgramManager functions.

iamalwaysuncomfortable · 2026-03-09T20:50:45Z

+    /// Check whether a specific program has been added.
+    ///
+    /// @param {string} name The program name.
+    /// @returns {boolean}
+    #[wasm_bindgen(js_name = "hasProgram")]
+    pub fn has_program(&self, name: &str) -> bool {
+        self.entries.contains_key(name)
+    }
+}


Similarly here

Suggested change

/// Check whether a specific program has been added.

///

/// @param {string} name The program name.

/// @returns {boolean}

#[wasm_bindgen(js_name = "hasProgram")]

pub fn has_program(&self, name: &str) -> bool {

self.entries.contains_key(name)

}

}

/// Check whether a specific program has been added.

///

/// @param {string} name The program name.

/// @returns {boolean}

#[wasm_bindgen(js_name = "contains")]

pub fn contains(&self, name: &str) -> bool {

self.entries.contains_key(name)

}

}

iamalwaysuncomfortable · 2026-03-09T20:52:50Z

+/// Extract source code from a JS value (plain string or object with `.source`).
+///
+/// Supports both legacy format (`"source code"`) and structured format
+/// (`{ source: "source code", ... }`).
+pub(crate) fn extract_source(value: &Option<wasm_bindgen::JsValue>) -> Option<String> {
+    let value = value.as_ref()?;
+    // Plain string format: "source code"
+    if let Some(source) = value.as_string() {
+        return Some(source);
+    }
+    // Structured object format: { source: "source code", ... }
+    Reflect::get(value, &"source".into()).ok().and_then(|v| v.as_string())
+}


Rename to program just to stick with existing conventions

Suggested change

/// Extract source code from a JS value (plain string or object with `.source`).

///

/// Supports both legacy format (`"source code"`) and structured format

/// (`{ source: "source code", ... }`).

pub(crate) fn extract_source(value: &Option<wasm_bindgen::JsValue>) -> Option<String> {

let value = value.as_ref()?;

// Plain string format: "source code"

if let Some(source) = value.as_string() {

return Some(source);

}

// Structured object format: { source: "source code", ... }

Reflect::get(value, &"source".into()).ok().and_then(|v| v.as_string())

}

/// Extract source code from a JS value (plain string or object with `.program`).

///

/// Supports both legacy format (`"source code"`) and structured format

/// (`{ program: "source code", ... }`).

pub(crate) fn extract_source(value: &Option<wasm_bindgen::JsValue>) -> Option<String> {

let value = value.as_ref()?;

// Plain string format: "source code"

if let Some(program) = value.as_string() {

return Some(program);

}

// Structured object format: { program: "source code", ... }

Reflect::get(value, &"source".into()).ok().and_then(|v| v.as_string())

}

iamalwaysuncomfortable · 2026-03-09T21:22:31Z

+    ///    (respecting transitive static imports via depth-first resolution).
+    /// 2. Inserts any pre-provided proving and verifying keys directly into
+    ///    the process, avoiding expensive on-demand key synthesis.
+    pub(crate) fn resolve_into(&self, process: &mut ProcessNative) -> Result<(), String> {


Maybe we rename this as load_programs?

Suggested change

pub(crate) fn resolve_into(&self, process: &mut ProcessNative) -> Result<(), String> {

pub(crate) fn load_programs(&self, process: &mut ProcessNative) -> Result<(), String> {

iamalwaysuncomfortable · 2026-03-09T21:24:26Z

+            if !process.contains_program(program.id()) {
+                log(&format!("Importing program: {name}"));
+                // Resolve transitive static imports first.
+                self.resolve_program_imports(process, &program)?;


It looks like this method might skip adding the program keys, perhaps we want to ensure that when we're calling this, any keys stored are also added?

iamalwaysuncomfortable · 2026-03-09T21:26:01Z

+    }
+
+    /// Recursively resolve a program's static imports in depth-first order.
+    fn resolve_program_imports(&self, process: &mut ProcessNative, program: &ProgramNative) -> Result<(), String> {


We might want to add any program proving and verifying keys here or consider just unifying this method with the method that calls it, because I believe as is, if programs get added here, it wouldn't add any of the proving or verifying keys.

…namic-dispatch branch

* Add WASM support for dynamic dispatch variant types (DynamicRecord, DynamicFuture, RecordWithDynamicID, ExternalRecordWithDynamicID) * fmt * fix(wasm): normalize dynamic dispatch type strings to match snarkVM serialization and add missing doc comments * add JS/TS fixture tests for dynamic dispatch variant types (record_dynamic, record_with_dynamic_id, external_record_with_dynamic_id)

…ture to match snarkVM (#1210)

* Handle Consensus V14 in deployments * Remove redundant .map_err in latest_stateroot call * Update getOrInitConsensusVersionTestHeights tests and doc comments to handle all currently active test version heights * Deallocate deployment cost tuple

* Add stringToField utility function for converting program and function names to fields for dynamic dispatch * Add exports of the stringToField utility in the js SDK

* Add the dynamic record type * Add JS side tests for the DynamicRecord type * Check visibility on record conversion roundtrip --------- Signed-off-by: Mike Turner <mike@provable.com>

* Handle Consensus V14 in deployments (#1230) * Handle Consensus V14 in deployments * Remove redundant .map_err in latest_stateroot call * Update getOrInitConsensusVersionTestHeights tests and doc comments to handle all currently active test version heights * Deallocate deployment cost tuple * Updates to make AMM tests works * Also handle upgrade * Address feedback * Cleanup * Cargo fmt lints * Revert .gitignore changes unrelated to this PR * Change core rev to the testnet 4.6.0 candidate * Bump version update to v0.9.18 * Remove unecessary dependencies in the workspace root --------- Co-authored-by: Mike Turner <mike@provable.com>

…ints with comprehensive test coverage

…patch imports in a single pass

… to imported programs

…ix truthy check and malformed JSDoc

… review nits

…eys across all execution paths

…stubs and review nits

…m to avoid key serialization, fix test stubs

…st coverage

marshacb changed the title ~~feat: Add dynamic dispatch import resolution across all WASM entry po…~~ feat: Add dynamic dispatch import resolution across WASM entry points Feb 26, 2026

marshacb force-pushed the feat/cam-dynamic-dispatch-imports branch from 4c5615d to ae46449 Compare February 26, 2026 06:13

vercel Bot deployed to Preview February 26, 2026 06:13 View deployment

marshacb force-pushed the feat/cam-dynamic-dispatch-imports branch from ae46449 to 15acf8e Compare February 26, 2026 06:36

vercel Bot deployed to Preview February 26, 2026 06:36 View deployment

marshacb force-pushed the feat/cam-dynamic-dispatch-imports branch from 15acf8e to af672cd Compare February 26, 2026 15:07

vercel Bot deployed to Preview February 26, 2026 15:07 View deployment

marshacb marked this pull request as ready for review February 26, 2026 15:34

cursor Bot reviewed Feb 26, 2026

View reviewed changes

marshacb requested a review from iamalwaysuncomfortable February 26, 2026 15:54

iamalwaysuncomfortable requested review from Copilot February 26, 2026 18:14

Copilot started reviewing on behalf of iamalwaysuncomfortable February 26, 2026 18:15 View session

Copilot AI reviewed Feb 26, 2026

View reviewed changes

Copilot AI reviewed Feb 26, 2026

marshacb changed the title ~~feat: Add dynamic dispatch import resolution across WASM entry points~~ [Feature]: Add dynamic dispatch import resolution across WASM entry points Mar 2, 2026

marshacb changed the title ~~[Feature]: Add dynamic dispatch import resolution across WASM entry points~~ [Feature] Add dynamic dispatch import resolution across WASM entry points Mar 2, 2026

vercel Bot deployed to Preview March 3, 2026 17:41 View deployment

vercel Bot deployed to Preview March 5, 2026 00:06 View deployment

iamalwaysuncomfortable requested a review from Copilot March 6, 2026 16:12

Copilot started reviewing on behalf of iamalwaysuncomfortable March 6, 2026 16:13 View session

Copilot AI reviewed Mar 6, 2026

View reviewed changes

vercel Bot deployed to Preview March 6, 2026 17:53 View deployment

iamalwaysuncomfortable requested changes Mar 9, 2026

View reviewed changes

iamalwaysuncomfortable mentioned this pull request Mar 9, 2026

[Feature] Return proving/verifying keys from execution methods #1227

Closed

iamalwaysuncomfortable force-pushed the feat/dynamic-dispatch branch from 0f51672 to ad9220e Compare March 11, 2026 19:33

marshacb force-pushed the feat/cam-dynamic-dispatch-imports branch from 54b34eb to b602f73 Compare March 11, 2026 21:11

vercel Bot deployed to Preview March 11, 2026 21:12 View deployment

marshacb force-pushed the feat/cam-dynamic-dispatch-imports branch from b602f73 to 50962b4 Compare March 11, 2026 21:18

marshacb requested a review from iamalwaysuncomfortable March 11, 2026 21:55

marshacb force-pushed the feat/cam-dynamic-dispatch-imports branch from 50962b4 to dcc4c69 Compare March 13, 2026 15:53

vercel Bot deployed to Preview March 13, 2026 15:54 View deployment

iamalwaysuncomfortable and others added 7 commits March 23, 2026 19:56

Change SnarkVM dep to latest rev on the corresponding SnarkVM:feat/dy…

a430afe

…namic-dispatch branch

fix: use dot-delimited type strings for dynamic.record and dynamic.fu…

367a7f2

…ture to match snarkVM (#1210)

Update SnarkVM rev for dynamic dispatch

25423eb

[Feature] Create stringToField utility (#1237)

02d19b4

* Add stringToField utility function for converting program and function names to fields for dynamic dispatch * Add exports of the stringToField utility in the js SDK

[Feature] Re-export dynamic record from SnarkVM (#1236)

a2d2ad7

* Add the dynamic record type * Add JS side tests for the DynamicRecord type * Check visibility on record conversion roundtrip --------- Signed-off-by: Mike Turner <mike@provable.com>

iamalwaysuncomfortable force-pushed the feat/dynamic-dispatch branch from e56f119 to a2d2ad7 Compare March 24, 2026 01:00

d0cd and others added 10 commits March 23, 2026 18:11

feat: Add dynamic dispatch import resolution across all WASM entry po…

30249f0

…ints with comprehensive test coverage

refactor: Unify resolve_imports to handle both static and dynamic dis…

180498b

…patch imports in a single pass

Add ProgramImportsBuilder for zero-cost proving/verifying key passing…

fa5d4c6

… to imported programs

Fix review feedback: guard phase-2 key insertion, remove debug log, f…

3ba2b7a

…ix truthy check and malformed JSDoc

Refactor ProgramEntry to store native types, add edition support, fix…

f8762c6

… review nits

Add KeyStore integration with auto-persist for top-level and import k…

2173787

…eys across all execution paths

Add key byte methods and toObject to ProgramImportsBuilder, fix test …

be5e5b7

…stubs and review nits

Fix review feedback: restore type: module, add programNames/getProgra…

953a731

…m to avoid key serialization, fix test stubs

Add transitive import resolution for dynamic dispatch targets with te…

3774d9d

…st coverage

marshacb force-pushed the feat/cam-dynamic-dispatch-imports branch from dcc4c69 to 3774d9d Compare March 24, 2026 14:27

vercel Bot deployed to Preview March 24, 2026 14:27 View deployment

iamalwaysuncomfortable force-pushed the feat/dynamic-dispatch branch from c0ac20e to a095833 Compare March 25, 2026 03:23

marshacb mentioned this pull request Mar 25, 2026

[Feature] Add ProgramImportsBuilder for zero-cost proving/verifying #1262

Merged

iamalwaysuncomfortable force-pushed the feat/dynamic-dispatch branch 2 times, most recently from 6c5a46a to 486b002 Compare March 26, 2026 07:29

Base automatically changed from feat/dynamic-dispatch to testnet March 26, 2026 07:33

iamalwaysuncomfortable force-pushed the testnet branch from b1382e8 to 3c76a12 Compare March 31, 2026 16:14

Base automatically changed from testnet to mainnet March 31, 2026 16:29

	entries: HashMap<String, ProgramEntry>,
	entries: HashMap<ProgramIDNative, ProgramEntry>,

	pub(crate) fn resolve_into(&self, process: &mut ProcessNative) -> Result<(), String> {
	pub(crate) fn load_programs(&self, process: &mut ProcessNative) -> Result<(), String> {

Uh oh!

Conversation

marshacb commented Feb 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Motivation

Test Plan

WASM tests (Rust):

SDK tests (TypeScript):

Manual testing:

Uh oh!

marshacb commented Feb 26, 2026

Uh oh!

cursor Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

iamalwaysuncomfortable left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

iamalwaysuncomfortable Mar 9, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

iamalwaysuncomfortable Mar 9, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

iamalwaysuncomfortable Mar 9, 2026

Choose a reason for hiding this comment

Uh oh!

iamalwaysuncomfortable Mar 9, 2026

Choose a reason for hiding this comment

Uh oh!

iamalwaysuncomfortable Mar 9, 2026

Choose a reason for hiding this comment

Uh oh!

iamalwaysuncomfortable Mar 9, 2026

Choose a reason for hiding this comment

Uh oh!

iamalwaysuncomfortable Mar 9, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

marshacb commented Feb 26, 2026 •

edited

Loading